Current opinion holds that carcinogenesis is a process of multiple genetic steps. Though the molecular nature of some of the steps that may be involved in neoplastic transformation is known, the manner in which individual steps are linked into a multistep process is obscure. A major impediment to identifying essential steps and determining their sequence in a tumor lineage stems from the difficulty of distinguishing those phenotypic/genotypic aberrations that are necessary for tumor formation (signal) from those that are not (noise). This project attempts to overcome this problem by analyzing the clonal cosegregation of tumorigenicity with specific phenotypic/genotypic properties among multiple cloned lines of liver epithelial cells. Cell lines are individually cloned from parental transformed populations that heterogeneously express both phenotypic/geneotypic properties and tumorigenicity; clonal lines produce tumors in from none to all of the syngeneic animals into which cells are transplanted, and many other cellular properties are equally variable. By identifying two or more properties that cosegregate within cloned lines, clonal analysis provides a powerful method to identify a mechanistic association among multiple cellular properties. Determining the temporal acquisition of properties that cosegregate with tumorigenicity delineates a tumor lineage, as well as identifies the individual steps that lead to the ability to form tumors. Delineation of a step-wise lineage, and establishment of cell populations representing different lineage stages, provides the cellular substrate with which to analyze the molecular mechanisms of each step. We study a cultured line of rat liver epithelial cells (WB-F344) which are phenotypically related to hepatocytes and produce hepatocellular carcinomas, hepatoblastomas and biliary adenocarcinomas when neoplastically transformed in vitro. We have identified a number of phenotypic properties that do not cosegregate with tumorigenicity among multiple clones, as well as several phenotypic properties that do cosegregate clonally with tumorigenicity. We have delineated one tumor lineage that includes several, apparently independent steps: nonrandom chromosomal changes, expression of transforming growth factor-alpha, and heightened expression of myc and ras oncogenes. During this project, we will continue studies to identify other tumor lineages and we will carry out a series of new studies to investigate the mechanisms that lead to the execution of individual steps and to the cellular responses that propel the affected cell through the process of transformation.